Identification of CDK1 As a Candidate Marker in Cutaneous Squamous Cell Carcinoma by Integrated Bioinformatics Analysis

Home , AP2S1, AP3B1, ATP5F1, ATP6V0B, ATP6V0D1, ATP6V0E1

478

Original Article

Identification of CDK1 as a candidate marker in cutaneous squamous cell carcinoma by integrated bioinformatics analysis

Si Qin1,2#, Yu Yang2,3#, Hao-Bin Zhang4, Xiao-Huan Zheng5, Hua-Run Li1,2, Ju Wen1,2

1Department of Dermatology, Guangdong Second Provincial General Hospital, Guangzhou, China; 2The Second School of Clinical Medicine, Southern Medical University, Guangzhou, China; 3Department of Urology, Peking University Shenzhen Hospital, Shenzhen, China; 4The Big Data Institute, Guangdong Create Environmental Technology Company Limited, Guangzhou, China; 5Nanhai District People’s Hospital, Foshan, China Contributions: (I) Conception and design: S Qin, Y Yang, HB Zhang; (II) Administrative support: J Wen; (III) Provision of study materials or patients: S Qin; (IV) Collection and assembly of data: XH Zheng, HR Li; (V) Data analysis and interpretation: S Qin, Y Yang, HB Zhang; (VI) Manuscript writing: All authors; (VII) Final approval of manuscript: All authors. #The authors contributed equally to this work. Correspondence to: Ju Wen. Department of Dermatology, Guangdong Second Provincial General Hospital, Guangzhou 510317, China. Email: [email protected].

Background: Cutaneous squamous cell carcinoma (cSCC) is a relatively common cancer that accounts for nearly 50% of non-melanoma skin cancer cases. However, the genotypes that are linked with poor prognosis and/or high relapse rates and pathogenic mechanisms of cSCC are not fully understood. To address these points, three gene expression datasets were analyzed to identify candidate biomarker genes in cSCC. Methods: The GSE117247, GSE32979, and GSE98767 datasets comprising a total of 32 cSCC samples and 31 normal skin tissue samples were obtained from the National Center for Biotechnology Information Gene Expression Omnibus database. Differentially expressed genes (DEGs) were identified and underwent pathway enrichment analyses with the Gene Ontology and Kyoto Encyclopedia of Genes and Genomes (KEGG). A putative DEG protein–protein interaction (PPI) network was also established that included hub genes. The expression of CDK1, MAD2L1, BUB1 ans CDC20 were examined in the study. Results: A total of 335 genes were identified, encompassing 219 found to be upregulated and 116 genes that were downregulated in cSCC, compared to normal tissue. Enriched functions of these DEGs were associated with Ephrin receptor signaling and cell division; cytosol, membrane, and extracellular exosomes; ATP-, poly(A) RNA-, and identical protein binding. We also established a PPI network comprising 332 nodes and identified KIF2C, CDC42, AURKA, MAD2L1, MYC, CDK1, FEN1, H2AFZ, BUB1, BUB1B, CKS2, CDC20, CCT2, ACTR2, ACTB, MAPK14, and HDAC1 as candidate hub genes. The expression of CDK1 are significantly higher in the cSCC tissues than that in normal skin. Conclusions: The DEGs identified in this study are potential therapeutic targets and biomarkers for cSCC. CDK1 is a gene closely related to the occurrence and development of cSCC, which may play an important role. Bioinformatics analysis shows that it is involved in the important pathway of the pathogenesis of cSCC, and may be recognized and applied as a new biomarker in the future diagnosis and treatment of cSCC.

Keywords: Cluster analysis; carcinoma; squamous cell; critical pathways; genetic association studies

Submitted Sep 18, 2020. Accepted for publication Nov 12, 2020. doi: 10.21037/tcr-20-2945 View this article at: http://dx.doi.org/10.21037/tcr-20-2945

© Translational Cancer Research. All rights reserved. Transl Cancer Res 2021;10(1):469-478 | http://dx.doi.org/10.21037/tcr-20-2945 470 Qin et al. Integrated bioinformatics analysis for cutaneous squamous cell carcinoma

Introduction of the study. We present the following article in accordance with Cancer is a major threat to human health and a global the MDAR reporting checklist (available at http://dx.doi. economic burden. The economic impact associated org/10.21037/tcr-20-2945). with anti-cancer therapies can largely affect treatment access and continuity, increasing the likelihood of poor prognoses (1). Cutaneous squamous cell carcinoma (cSCC) Methods is a relatively common cancer that accounts for nearly Microarray analysis and identification of DEGs 50% of non-melanoma skin cancer cases (2). In addition to environmental exposure, a variety factors such as aging, The three microarray datasets of primary cSCC and normal human papilloma virus infection, immunosuppression, tissues (GSE117247, GSE32979, and GSE98767) were genetic factors and in particular, ultraviolet radiation obtained from the NCBI GEO database. The GSE117247 are all associated with increased cSCC risk. At present data was based on the Affymetrix Human Genome U95 the preferred treatment of cSCC is radical resection. v2 array (GPL8300 platform), derived from eight primary Although most cSCC patients who undergo cruative cSCC and nine normal skin samples. The GSE32979 surgery exhibit favorable prognoses, a minority develop microarray data were obtained using the Illumina Human-6 metastasis or recurrence that is potentially fatal. For v2.0 expression bead chip (GPL6102 platform) from 15 patients with advanced cSCC, due to the lack of large- cSCC and 13 normal skin samples. GSE98767 data were scale clinical trials, clinicians have to rely on the efficacy obtained using the Illumina Human HT-12 v4.0 expression evidence of other tumor types (such as head and neck bead chip (GPL10558 platform) from nine primary cSCC mucosal squamous cell carcinoma) to guide treatment and nine normal tissue samples. The integration and plan. Systemic therapy, like platinum chemotherapy, were analysis of all three datasets was performed with GEO2R, used in advanced cSCC patients population (3,4). Genetic a web-based analysis tool [https://www.ncbi.nlm.nih. alterations have shown to play a key role in the resistance gov/geo/geo2r/?acc=GSE98767], and saved.. Genes to anticancer drugs in oncology patients, demonstrating differentially expressed between cSCC and normal tissues an urgent need to identify key genes that can serve as were screened according to a standard P value of <0.05, and therapeutic targets and biomarkers for monitoring the fold change (FC) in gene expression >0 (|logFC| >0). The response to treatment (5). Recently, EGFR inhibitors, DEGs were categorised into Venn diagrams using a web- cemiplimab and pembrolizumab which aim at biomarkers based application (http://bioinformatics.psb.ugent.be/ were used for patients with advanced cSCC. webtools/Venn/). The study was conducted in accordance The development of microarray and high-throughput with the Declaration of Helsinki (as revised in 2013). sequencing technologies has facilitated the identification of biomarkers related to tumor occurrence and progression GO and KEGG pathway enrichment analyses (6-8). Unlike head and neck squamous cell carcinoma, relatively few marker genes have been identified in cSCC The online gene function annotation tool, Database for (9,10). Annotation, Visualization, and Integrated Discovery In recent research, Wei et al. (11) suggested that EGR3 (DAVID), (https://david.ncifcrf.gov/), was used to analyze had the potential to help in the diagnosis and predictive the functions and pathway enrichment of candidate DEGs. treatments of CSCC by bioinformatics, but the samples in The cut-off for calculating the degree of enrichment was the study are not from the same organization or the same P<0.05. GEO profiles. To address this deficiency, bioinformatic analysis was performed in our study for providing more PPI networks, clustering, and identification of hub genes reliable research evidence for the early diagnosis and prevention of cSCC the development and progression. In Proteins corresponding to DEGs and a PPI network our study, we used the same tissue-derived specimens, and diagram were obtained with STRING (https://string- compare the different genes in the same GEO profile which db.org), an online tool for identifying functional protein helped to avoid partial bias. In addition, we validated the association networks. Candidate genes were evaluated results of bioinformatics in order to enhance the reliability with the Cytoscape (Version 3.7.2) software to analyze

GSE117247 GSE32979 GSE117247 GSE32979 for 3–5 min, followed by rinsing with running water for 5–10 min to stop the color development. Then, the tissue 388 416 sections were counterstained with hematoxylin for 1 min 566 2334 663 2883 and differentiated with 0.5% hydrochloric alcohol for 219 116 2 seconds. The sections were dehydrated and mounted 236 985 161 822 for microscopic examination at room temperature. The signal was visualized with Light microscopy (Olympus 3236 3523 BX43; Olympus Corporation) (magnification, ×200) was

GSE98767 GSE98767 performed, and the presence of brown granules in the cytoplasm was considered a positive signal. up-regulation genes down-regulation genes

Figure 1 DEGs of three GEO profiles. Detailed explanation: 219 Results upregulated genes and 116 downregulated genes were filtered from three GEO profiles with LogFC >0 and LogFC <0. DEGs, Identification of DEGs in cSCC differentially expressed genes; GEO, Gene Expression Omnibus. Gene expression profiles of 32 cSCC and 31 normal skin tissue samples were compared in this study. GEO2R interactions between genes. All analyzed genes and modules analysis yielded 2749, 8062, and 8472 identified DEGs in of PPI network underwent further screening with the GSE117247, GSE32979, GSE98767 datasets, respectively. MCODE plug-in [degree cut off =2 (12), node score cut off Among the 756 genes co-expressed across the three =0.2, k-core =2, max. Depth =100]. datasets, 219 were shown to be upregulated (logFC >0) and 116 were downregulated (logFC <0) in cSCC compared to normal tissue (Figure 1 and Table 1). Statistical analysis

Data were analyzed with the paired-samples t test using Analysis of GO terms for DEGs in cSCC Prism 6 software (GraphPad, La Jolla, CA, USA). P<0.05 was considered statistically significant. GO analysis of the DEGs was undertaken using P<0.05 as the cut off. Under the biological process (BP) domain, DEGs were mainly enriched in the Ephrin receptor Immunohistochemical (IHC) analysis signaling pathway and cell division. The top three processes All samples (including normal tissues adjacent to carcinoma under Cellular Components (CC) domain were cytosol, and cSCC tissues) obtained from the 20 cSCC patients membrane, and extracellular exosome. Protein-, ATP, were fixed in 10% formalin for 24 h at room temperature poly(A) RNA-, and identical protein binding were the and embedded in paraffin. The paraffin-embedded tissues highest ranked processes under the Molecular function (MF) were serially cut into 5μm sections and heated in an oven domain (Figure 2). at 60 for 2 h. The tissue sections were dewaxed, antigen- repaired according to high temperature and high pressure. ℃ KEGG pathway enrichment analysis After antigen repair, the sections were rinsed with PBS for 3 times, and incubated with Cdc2 p34 (200 μg/mL; 1:200; Signaling pathways associated with the DEGs were cat. no. sc-54; Santa Cruz), BUB1(B-3) (200 μg/mL; 1:200; examined, with P<0.05 as the cut off. Identified candidate cat. no. sc-365685; Santa Cruz), MAD2(C-10) (200 μg/mL; pathways, linked to DEGs were shown to be involved in 1:200; cat. no. sc-374131; Santa Cruz) for 18 h at 4 , various BPs, including infection (epithelial cell signaling in respectively. The sections in incubation box was rinsed with Helicobacter pylori infection, pathogenic escherichia coli ℃ PBS 3 times for 5 minutes. The excess PBS was removed, infection, hepatitis C), cell proliferation and death (cell and the m-IgGκ BP-HRP (cat. no. sc-516102, Santa Cruz) cycle, oocyte meiosis), immunity (Gamma R-mediated was added before incubation at 37 for 30 min. And then phagocytosis), and metabolism (lipolysis regulation in the sections were rinsed with PBS for 3 times. DAB Color- adipocytes) (Table 2). The data collectively suggests that the ℃ substrate solution (cat. no. K5007; DAKO) was added main pathways involved in the development of cSCC are

© Translational Cancer Research. All rights reserved. Transl Cancer Res 2021;10(1):469-478 | http://dx.doi.org/10.21037/tcr-20-2945 472 Qin et al. Integrated bioinformatics analysis for cutaneous squamous cell carcinoma

Table 1 Genes differentially expressed between cSCC and normal tissue DEGs Gene names

Down-regulated ABR, MAOA, GNA11, ARAP1, KRT15, ITSN1, DCT, KCTD2, CRTAP, ABCA6, LGALS3, LPCAT4, KIAA0895, MAN2B2, genes MBP, CDH19, CTIF, ESRRG, CNGA1, CHKB, MAPRE3, GCHFR, ATG9A, CAMK2G, LSS, ID4, IQGAP1, GSTM3, ACVR1B, AMT, AHNAK2, CDK10, ITPR3, GHR, DNM1, ALDH6A1, DIXDC1, ARHGAP12, MXRA5, DAG1, MIA, CCS, CTNNBIP1, HMGN2, ASMTL, MZF1, ABCA2, FXYD1, FASTK, EFS, ASPA, AGTR1, EPHX2, BPTF, DST, ABCA3, N4BP2L2, BCL2, ZBTB20, FAM193A, MAN2C1, EEF1D, CRELD1, GPC3, AKAP8, ABCA5, MACF1, CAPN3, DIAPH2, CRYAB, GSN, HOXA10, SLIT3, ITPR2, GOSR1, EZH1, CIC, HTT, FAM172A, MON2, ARMCX6, ARHGEF10, ATG14, B3GAT3, AHDC1, MITF, GABBR1, FYN, BARD1, CBX7, CYB5R3, ICA1, LAMB2, NBAS, FABP7, ARHGEF7, MAN2A2, MAU2, DDX42, FBLN1, BMP4, ESD, LAMA5, TSPOAP1, ECM2, LYRM9, ACTR1B, SLF2, BAP1, FAM189A2, CADM1, DNAJB12, HOXC6, DKFZP586I1420, ERCC2

Up-regulated ATP5F1, LMAN1, LMNB2, FAM98A, FOXM1, CDK1, GCLM, ABCF2, GSTP1, KIF14, MAD2L1, EIF4A3, GOT2, MANF, genes EIF4E2, C1QBP, CASP1, GCH1, MELK, GPX2, NDC80, AP2S1, MYC, CCNA2, ISCA1, ALAS1, CRLF3, AMPD3, CKS2, BCR, CCNB2, MPHOSPH6, HIST1H2AC, DDX21, AZIN1, CBX3, NCK1, PUDP, CSTB, GTF2E2, CCNF, HDAC1, CNIH1, BLZF1, EIF3I, CASP4, DYNLT3, CSNK2A1, TYMP, ABLIM1, ACTR2, ATOX1, BID, COX5A, ATP6V0D1, GNAI3, FCGR2B, ABCG1, LYPLA1, NAPG, HMOX2, ATP6V0B, AKIRIN2, KIF11, CCT2, ARF4, MED8, DSC2, MID1, DBF4, AHCY, MAPK14, ARPP19, MPZL1, NOP16, CAP1, CDC42, EREG, CDC20, GART, DNAJB6, H2AFZ, CCNT2, BUB1, DFNA5, HNRNPC, DUSP6, MRPL3, MAPK6, EIF5, TPGS2, DPH2, FEN1, HPRT1, HIF1A, ACTB, CACYBP, CSRP2, CUX1, AIM2, HDGF, ATP1B3, E2F3, GARS, H2AFY, HSD17B2, CSTF1, BAK1, PUM3, KPNA2, MRPS12, ACOT7, KIF2A, GLA, ARPC3, APOBEC3B, AURKA, AP3B1, C6orf62, FAM189B, IFI16, CEBPD, DDX39A, ERCC1, DUSP14, EIF2B2, ARFGEF2, UBE2K, ARPC1A, HIST1H2BE, FGFBP1, CDH3, EI24, COPB2, GGH, HMOX1, CTNS, GADD45A, GTF2A2, COPE, FSCN1, CFL1, GBP1, ATP5J, ACVR1, EIF4EBP1, GSPT1, GGCT, HIST1H2BK, DLGAP5, ENO1, GOSR2, ADAM17, EIF4H, CTSA, APEX2, DNM1L, CDC42EP1, KIAA0101, BCL10, CDA, EPHB2, CKAP4, AGFG1, HMGB3, HOMER3, GK, ASCC3, EPHA2, CYFIP1, DYNLT1, GNG5, MTHFD2, CSK, MORF4L2, BZW1, KYNU, ETV4, CCL20, DNAJA1, KIF2C, CYB561D2, ADAM10, GNA15, CLCA2, AK2, ILF3, CSNK1A1, CCNC, DCTN6, BDH1, CD24, CSTF2, DNMT1, CAPRIN1, HOMER1, HK2, NCBP1, ADK, CA2, GRN, FGFR1OP, FTSJ1, FLAD1, DSG3, BUB1B, FMO2, ALG3, EXOG, ATP12A, GIPC1, ATP6V1D, EIF4G2, ATP2A2, CASK, ATP6V0E1, AMD1, CTSC cSCC, Cutaneous Squamous Cell Carcinoma; DEGs, Differentially Expressed Genes.

those associated with metabolism. We also analyzed the top 30 hub genes with the Betweennes, MCC, MNC, Degree, EPC, BottleNeck, EcCentricity, Closenes, Radiality, and stress methods. The PPI network and modular analyses 17 hub genes identified by these nine statistical methods An integrated analysis of the 335 candidate genes was were KIF2C, CDC42, AURKA, MAD2L1, MYC, CDK1, performed using the STRING database, generating a PPI FEN1, H2AFZ, BUB1, BUB1B, CKS2, CDC20, CCT2, network diagram which excluded 332 nodes and 1,307 ACTR2, ACTB, MAPK14, and HDAC1. These DEGs are edges. Pathway analyses using GO and KEGG (Figure 3) critically involved in the pathogenesis and development of revealed that the DEGs were mainly enriched in pathways cSCC. related to Epithelial cell signaling in H. pylori infection and the cell cycle. Analysis of the DEGs with Cytoscape Expression of CDK1 in cSCC software and the MCODE plug-in was performed. The top two functional modules are shown in Figure 4. Module Based on the above results, CDK1 expression in cSCC was 1 exhibited 20 nodes and 185 edges, whilst Module 2 analyzed in tumor and normal tissues via IHC analysis. showed 55 nodes and 391 edges. GO and KEGG pathway The results are shown in Figure 5A,B, where cSCC tissue is enrichment analyses of the DEGs in both modules displayed indicated by the presence of positively stained brown nuclei. 20 upregulated and no downregulated genes in Module 1 CDK1 expression was higher in cSCC compared to normal (Figure 4A). However, Module 2 (Figure 4B), contained 47 skin tissue (Figure 5C,D), consistent with results obtained by upregulated genes, and 8 that were downregulated. The integrating the GEO profiles. These results suggested that genes in both modules were mainly related to cell cycle CDK1 may serve as a new biomarker for the diagnosis and functions (Table 3). prognosis of cSCC. The expression of MAD2L1 and BUB1

−LOG (P value)

0.00 5.00 10.00 15.00 20.00 25.00 protein binding ATP binding poly(A) RNA binding identical protein binding ATPase activity protein homodimerization activity protein kinase binding hydrogen ion transmembrane transporter activity enzyme binding protein serine/threonine kinase activity cytosol membrane extracellular exosome cytoplasm nucleoplasm protein complex myelin sheath cell-cell adherens junction focal adhesion mitochondrion ephrin receptor signaling pathway cell division mitotic nuclear division regulation of translational initiation intra-Golgi vesicle-mediated transport cell proliferation toxin transport G1/S transition of mitotic cell cycle negative regulation of extrinsic apoptotic signaling pathway release of cytochrome C from mitochondria

Molecular FunctionI Celluar Componentl Biological Process

Figure 2 GO enrichment of DEGs. Detailed explanation: Go enrichment of DEGs in 3 functional groups: molecular function (blue), cellular components (orange) or biological processes (grey). GO, Gene Ontology; DEGs, differentially expressed genes.

Table 2 Functional enrichment analysis of signaling pathways associated with differentially expressed genes in cSCC Pathway Gene count P value Genes

Cell cycle 12 3.29E-03 MAD2L1, BUB1, BUB1B, CDC20, CDK1, E2F3, CCNB2, HDAC1, DBF4, CCNA2, GADD45A, MYC

Epithelial cell signaling in 9 1.97E-03 CDC42, ATP6V0E1, ADAM10, MAPK14, ADAM17, ATP6V0D1, CSK, Helicobacter pylori infection ATP6V1D, ATP6V0B

Oocyte meiosis 9 3.29E-02 CDK1, MAD2L1, CCNB2, CAMK2G, BUB1, CDC20, AURKA, ITPR3, ITPR2

Fc gamma R-mediated 8 4.29E-02 ARPC1A, CDC42, DNM1L, ARPC3, FCGR2B, GSN, CFL1, DNM1 phagocytosis

Pathogenic Escherichia coli 6 4.30E-02 ARPC1A, ACTB, CDC42, ARPC3, FYN, NCK1 infection cSCC, cutaneous squamous cell carcinoma.

protein in cSCC showed no significant difference compare to screen 756 genes in three GEO expression profiles, with that in normal tissues. including 219 that were shown to be upregulated and 116 that were downregulated in cSCC, compared to normal tissue. In the BP domain of the GO analysis, most DEGs Discussion were enriched in Ephrin receptor signaling and cell In this study, we used integrated bioinformatics analysis division, which have been implicated in angiogenesis, stem

© Translational Cancer Research. All rights reserved. Transl Cancer Res 2021;10(1):469-478 | http://dx.doi.org/10.21037/tcr-20-2945 474 Qin et al. Integrated bioinformatics analysis for cutaneous squamous cell carcinoma

−LOG (P value) 0 0.5 1 1.5 2 2.5 3 3.5 4

hsa05120: Epithelial cell signaling in Helicobacter pylori infection

hsa04110: Cell cycle

hsa04114: Oocyte meiosis

hsa04666: Fc gamma R-mediated phagocytosis

hsa05130: Pathogenic Escherichia coli infection

Figure 3 KEGG pathway enrichment analysis of DEGs. Detailed explanation: the DEGs were mainly enriched in Epithelial cell signaling in Helicobacter pylori infection and Cell cycle pathways. KEGG, Kyoto Encyclopedia of Genes and Genomes; DEGs, differentially expressed genes.

A B AGTR1 AGFG1

IQGAP1 GSN

AP2S1 CYFIP1 ITSN1 DNM1 CYB5R3 GLA COPE KPNA2 BUB1B ACTR2 CDC42 EPHB2 CTSA COPB2 KIF2C CAP1 FEN1 KIF2A DCTN6 CCNA2 FSCN1 GGH NDC80 CCNB2 CTSC GRN KIF2C CCNF KIF11 MAD2L1 ARF4 CKS2 MYC LMNB2 CCT2 CDK1 BUB1 CDC20 KIF11 FOXM1 NDC80 BUB1B CDK1 KIF14 MAD2L1 CCNA2 AURKA FEN1 KPNA2 CCNB2 DLGAP5 E2F3 FOXM1 DLGAP5 XACDC20 MELK KIAA0101 CKS2 MELK H2AFZ CCNF DBF4 KIF14 BARD1 BUB1 AURKA CBX3 DNMT1

H2AFY

Figure 4 Hub genes in the PPI network. Detailed explanation: Hub genes in Module 1 (A) and Module 2 (B). Specific PPIs are represented as lines. PPI, protein-protein interaction. cell differentiation, and cancer. Cytosol and protein binding Cytoscape MCODE were found to comprise 20 and 55 were enriched in the CC and MF domains, respectively. genes, respectively. Cell division was the most significant KEGG analysis also revealed Epithelial cell signaling in BP in Modules 1 and 2. H. pylori infection, cell cycle, and oocyte meiosis as the Most cancers are associated with genetic mutations highest ranking pathways associated with the DEGs. In involved in cell cycle regulation or aberrant activities the PPI network, two co-expression clusters analyzed with of cyclin-dependent kinases (CDK), cyclins, and CDK

Table 3 GO and pathway enrichment analysis of module 1 and 2 genes

Term Description Count P value

Module 1

GO:0051301 (Biological process) Cell division 14 2.7×1018

GO:0000776 (Cellular component) Kinetochore 5 1.3×106

GO:0019901 (Molecular function) Protein kinase binding 6 4.8×105

hsa:04110 (KEGG pathway) Cell cycle 7 6.0×109

Module 2

GO:0051301 (Biological process) Cell division 17 6.7×1015

GO:0005829 (Cellular component) Cytosol 29 2.2×108

GO:0005515 (Molecular function) Protein binding 46 5.0×107

hsa:04110 (KEGG pathway) Cell cycle 10 2.2×109 GO, Gene Ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes.

200 μm 100 μm

A B

200 μm 100 μm

C D

Figure 5 CDK1 expression in skin and cSCC. Immunohistochemical staining of CDK1 in cSCC tissues (A,B) and normal samples (C,D). cSCC, cutaneous squamous cell carcinoma. inhibitors (8-10), which are normally required in regulating cancer (16), breast cancer (17), and malignant pleural cell growth. CDK1 encodes a Ser/Thr kinase that plays mesothelioma (18). an essential part in the eukaryotic cell cycle and was Mitotic arrest defect 2 ligand 1 (MAD2L1) encodes a widely found to be overexpressed in oral SCC (13), liver cell cycle checkpoint protein that modulates anaphase- cell carcinoma (14), epithelial ovarian cancer (15), and promoting complex/cyclosome (APC/C) to ensure the breast cancer (15). In contrast, gene silencing of CDK1 correct arrangement of chromosomes on the metaphase has exhibited therapeutic potential in epithelial ovarian plate during cell division (19). MAD2L1 expression has

© Translational Cancer Research. All rights reserved. Transl Cancer Res 2021;10(1):469-478 | http://dx.doi.org/10.21037/tcr-20-2945 476 Qin et al. Integrated bioinformatics analysis for cutaneous squamous cell carcinoma recently been linked to tumor stage and prognosis (20). originated from highly heterogeneous patient populations, Additionally, high expression of MAD2L1 has been which may have affected the gene expression profiles. observed in breast cancer (21), gastric cancer (22), and other Nonetheless, 335 candidate DEGs were obtained by GEO malignancies. datasets and bioinformatics analysis. A total of 17 hub genes Mutation of Cell Division Cycle protein (CDC)20 leads were found to be mainly related to the epithelial signal to aberrant mitotic arrest and chromosome segregation (23). pathway, cell cycle and oocyte meiosis in cSCC and the high The activation of APC/C-CDC20 enhances the sensitivity expression of CDK1 in CSCC samples was verified. Most of cancer cells to radiotherapy and chemotherapy (24,25). importantly, these findings can improve our understanding CDC20 overexpression was recently reported in gastric of the causes and processes of cSCC. The identification cancer, colorectal cancer, hepatocellular carcinoma, cervical of potential genes and pathways in CSCC could aid in cancer, and bladder cancer. Moreover, CD20 overexpression identifying effective targeted therapies and contribute to its was highly correlated with poor prognoses (26-29). clinical management. Conversely, downregulation of CDC20 slowed growth and colony formation in small and non-small cell lung carcinoma Acknowledgments cells (30). The budding uninhibited by benzimidazoles 1 homolog Funding: This work was supported by a grant from the beta (BUB1B) gene encodes a kinase that is associated with Project of Administration of Traditional Chinese Medicine cell cycle checkpoints and chromosome segregation (31). of Guangdong, China (no. 20181011). The authors thank BUB1B binds to CDC20 during the G2 phase of mitosis Dr. Chaohu Wang for his expertise and guidance during the to inhibit APC/C activity and cell cycle progression (32). study. The upregulated BUB1B level was found to be highly correlated with unfavorable prognosis in glioblastoma (33) Footnote and hepatocellular carcinoma (34). Furthermore, its overexpression is thought to enhance the resistance of Reporting Checklist: The authors have completed the glioblastoma to radiotherapy (35). MDAR reporting checklist. Available at http://dx.doi. To date, few studies have undertaken detailed org/10.21037/tcr-20-2945 investigations into the precise role of CDK1 in the development and progression of cSCC. Our findings Conflicts of Interest: All authors have completed the ICMJE demonstrate that CDK1 was highly expressed in cSCC, uniform disclosure form (available at http://dx.doi. providing a sound rationale for future studies in rapidly org/10.21037/tcr-20-2945). The authors have no conflicts establishing the role and mechanism of CDK1 in cSCC. of interest to declare. The development of cSCC is complex, involving many histopathological and molecular changes. Therefore, Ethical Statement: The authors are accountable for all the characterization of cSCC will require additional aspects of the work in ensuring that questions related investigations beyond genetic profiling. However, the to the accuracy or integrity of any part of the work are analysis of gene expression may serve as a predictive tool appropriately investigated and resolved. The study was in tumor development and progression, and may aid in conducted in accordance with the Declaration of Helsinki (as identifying effective therapies in cSCC. revised in 2013). In summary, our findings have led to a better understanding of the mechanism, potential MFs and Open Access Statement: This is an Open Access article characteristics of cSCC via gene expression profiling and distributed in accordance with the Creative Commons bioinformatics analysis. These candidate genes can serve as Attribution-NonCommercial-NoDerivs 4.0 International biomarkers for the diagnosis, treatment and evaluation of License (CC BY-NC-ND 4.0), which permits the non- cSCC. commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the Conclusions formal publication through the relevant DOI and the license). It is worth noting that the samples analyzed in this study See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References 14. Zhao J, Han SX, Ma JL, et al. The role of CDK1 in apoptin-induced apoptosis in hepatocellular carcinoma 1. Landwehr MS, Watson SE, Macpherson CF, et al. cells. Oncol Rep 2013;30:253-9. The cost of cancer: a retrospective analysis of the 15. Chae SW, Sohn JH, Kim DH, et al. Overexpressions of financial impact of cancer on young adults. Cancer Med Cyclin B1, cdc2, p16 and p53 in human breast cancer: the 2016;5:863-70. clinicopathologic correlations and prognostic implications. 2. Rogers HW, Weinstock MA, Feldman SR, et al. Incidence Yonsei Med J 2011;52:445-53. Estimate of Nonmelanoma Skin Cancer (Keratinocyte 16. Xi Q, Huang M, Wang Y, et al. The expression of Carcinomas) in the U.S. Population, 2012. JAMA CDK1 is associated with proliferation and can be a Dermatol 2015;151:1081-6. prognostic factor in epithelial ovarian cancer. Tumour 3. Stratigos A, Garbe C, Lebbe C, et al. Diagnosis and Biol 2015;36:4939-48. treatment of invasive squamous cell carcinoma of the skin: 17. Liu Y, Zhu YH, Mao CQ, et al. Triple negative breast European consensus-based interdisciplinary guideline. Eur cancer therapy with CDK1 siRNA delivered by cationic J Cancer 2015;51:1989-2007. lipid assisted PEG-PLA nanoparticles. J Control Release 4. Keeping S, Xu Y, Chen CI, et al. Comparative efficacy of 2014;192:114-21. cemiplimab versus other systemic treatments for advanced 18. Linton A, Cheng YY, Griggs K, et al. An RNAi-based cutaneous squamous cell carcinoma. Future Oncol 2020. screen reveals PLK1, CDK1 and NDC80 as potential [Epub ahead of print]. therapeutic targets in malignant pleural mesothelioma. Br 5. Sordillo JE, Kraft P, Wu AC, et al. Quantifying the J Cancer 2018;118:e13. Polygenic Contribution to Cutaneous Squamous Cell 19. Cheng Y, Li K, Diao D, et al. Expression of KIAA0101 Carcinoma Risk. J Invest Dermatol 2018;138:1507-10. protein is associated with poor survival of esophageal 6. Chen X, Zhang FH, Chen QE, et al. The clinical cancer patients and resistance to cisplatin treatment in significance of CDK1 expression in oral squamous cell vitro. Lab Invest 2013;93:1276-87. carcinoma. Med Oral Patol Oral Cir Bucal 2015;20:e7-12. 20. Shi YX, Zhu T, Zou T, et al. Prognostic and predictive 7. Bonelli P, Tuccillo FM, Borrelli A, et al. CDK/CCN and values of CDK1 and MAD2L1 in lung adenocarcinoma. CDKI alterations for cancer prognosis and therapeutic Oncotarget 2016;7:85235-43. predictivity. Biomed Res Int 2014;2014:361020. 21. Yuan B, Xu Y, Woo JH, et al. Increased expression of 8. Yamamoto E, Kohama G, Sunakawa H, et al. Mode mitotic checkpoint genes in breast cancer cells with of invasion, bleomycin sensitivity, and clinical course chromosomal instability. Clin Cancer Res 2006;12:405-10. in squamous cell carcinoma of the oral cavity. Cancer 22. Kim HS, Park KH, Kim SA, et al. Frequent mutations 1983;51:2175-80. of human Mad2, but not Bub1, in gastric cancers 9. Malumbres M. Physiological relevance of cell cycle cause defective mitotic spindle checkpoint. Mutat Res kinases. Physiol Rev 2011;91:973-1007. 2005;578:187-201. 10. Shintani S, Mihara M, Nakahara Y, et al. Expression 23. Hartwell LH, Culotti J, Reid B. Genetic control of the of cell cycle control proteins in normal epithelium, cell-division cycle in yeast. I. Detection of mutants. Proc premalignant and malignant lesions of oral cavity. Oral Natl Acad Sci U S A 1970;66:352-9. Oncol 2002;38:235-43. 24. Wan L, Tan M, Yang J, et al. APC(Cdc20) suppresses 11. Wei W, Chen Y, Xu J, et al. Identification of Biomarker for apoptosis through targeting Bim for ubiquitination and Cutaneous Squamous Cell Carcinoma Using Microarray destruction. Dev Cell 2014;29:377-91. Data Analysis. J Cancer 2018;9:400-6. 25. Taniguchi K, Momiyama N, Ueda M, et al. Targeting of 12. Feng H, Gu ZY, Li Q, et al. Identification of significant CDC20 via small interfering RNA causes enhancement genes with poor prognosis in ovarian cancer via of the cytotoxicity of chemoradiation. Anticancer Res bioinformatical analysis. J Ovarian Res 2019;12:35. 2008;28:1559-63. 13. Chang JT, Wang HM, Chang KW, et al. Identification 26. Li J, Gao JZ, Du JL, et al. Increased CDC20 expression of differentially expressed genes in oral squamous cell is associated with development and progression of carcinoma (OSCC): overexpression of NPM, CDK1 and hepatocellular carcinoma. Int J Oncol 2014;45:1547-55. NDRG1 and underexpression of CHES1. Int J Cancer 27. Kim Y, Choi JW, Lee JH, et al. MAD2 and CDC20 are 2005;114:942-9. upregulated in high-grade squamous intraepithelial lesions

© Translational Cancer Research. All rights reserved. Transl Cancer Res 2021;10(1):469-478 | http://dx.doi.org/10.21037/tcr-20-2945 478 Qin et al. Integrated bioinformatics analysis for cutaneous squamous cell carcinoma

and squamous cell carcinomas of the uterine cervix. Int J BubR1 N terminus acts as a soluble inhibitor of cyclin B Gynecol Pathol 2014;33:517-23. degradation by APC/C(Cdc20) in interphase. Dev Cell 28. Ding ZY, Wu HR, Zhang JM, et al. Expression 2009;16:118-31. characteristics of CDC20 in gastric cancer and its 33. Lee E, Pain M, Wang H, et al. Sensitivity to BUB1B correlation with poor prognosis. Int J Clin Exp Pathol Inhibition Defines an Alternative Classification of 2014;7:722-7. Glioblastoma. Cancer Res 2017;77:5518-29. 29. Choi JW, Kim Y, Lee JH, et al. High expression of 34. Zhuang L, Yang Z, Meng Z. Upregulation of BUB1B, spindle assembly checkpoint proteins CDC20 and MAD2 CCNB1, CDC7, CDC20, and MCM3 in Tumor Tissues is associated with poor prognosis in urothelial bladder Predicted Worse Overall Survival and Disease-Free cancer. Virchows Arch 2013;463:681-7. Survival in Hepatocellular Carcinoma Patients. Biomed 30. Kidokoro T, Tanikawa C, Furukawa Y, et al. CDC20, a Res Int 2018;2018:7897346. potential cancer therapeutic target, is negatively regulated 35. Ma Q, Liu Y, Shang L, et al. The FOXM1/BUB1B by p53. Oncogene 2008;27:1562-71. signaling pathway is essential for the tumorigenicity 31. Guo Y, Kim C, Ahmad S, et al. CENP-E--dependent and radioresistance of glioblastoma. Oncol Rep BubR1 autophosphorylation enhances chromosome 2017;38:3367-75. alignment and the mitotic checkpoint. J Cell Biol 2012;198:205-17. (English Language Editor: E Tan) 32. Malureanu LA, Jeganathan KB, Hamada M, et al.

Cite this article as: Qin S, Yang Y, Zhang HB, Zheng XH, Li HR, Wen J. Identification of CDK1 as a candidate marker in cutaneous squamous cell carcinoma by integrated bioinformatics analysis. Transl Cancer Res 2021;10(1):469-478. doi: 10.21037/tcr- 20-2945